RFR: 8342642: Class loading failure due to archived map issue in ModuleLoaderMap.Mapper [v2]

[Jiangli Zhou]

On Fri, 25 Oct 2024 20:11:36 GMT, Jiangli Zhou <jiangli at openjdk.org> wrote:

>> Please review the fix that uses String type for the mapped value in ModuleLoaderMap.Mapper map (Map<String, String>). Please see details in https://bugs.openjdk.org/browse/JDK-8342642, thanks.
>
> Jiangli Zhou has updated the pull request incrementally with 109 additional commits since the last revision:
> 
>  - Reflect review feedback:
>    - Rename [APP|PLATFORM]_LOADER_INDEX to [APP|PLATFORM]_LOADER_NAME.
>    
>    - Update loader map index related comments.
>    
>    - Change to use '.equals()' (from '==') when comparing loader map values.
>    
>    Runs on Debian 6.9.10
>    
>    With '=='
>     Performance counter stats for 'images/jdk/bin/java -cp /.../home/jianglizhou/tests/hw.jar HelloWorld' (100 runs):
>    
>                133.08 msec task-clock:u                     #    1.181 CPUs utilized               ( +-  0.29% )
>                     0      context-switches:u               #    0.000 /sec
>                     0      cpu-migrations:u                 #    0.000 /sec
>                 5,940      page-faults:u                    #   44.636 K/sec                       ( +-  0.02% )
>           139,151,067      cycles:u                         #    1.046 GHz                         ( +-  0.24% )
>           158,618,665      instructions:u                   #    1.14  insn per cycle              ( +-  0.00% )
>            27,343,220      branches:u                       #  205.471 M/sec                       ( +-  0.00% )
>               746,103      branch-misses:u                  #    2.73% of all branches             ( +-  0.63% )
>    
>              0.112697 +- 0.000386 seconds time elapsed  ( +-  0.34% ).
>    
>    With '.equals()':
>     Performance counter stats for 'images/jdk/bin/java -cp /usr/local/google/home/jianglizhou/tests/hw.jar HelloWorld' (100 runs):
>    
>                133.26 msec task-clock:u                     #    1.186 CPUs utilized               ( +-  0.34% )
>                     0      context-switches:u               #    0.000 /sec
>                     0      cpu-migrations:u                 #    0.000 /sec
>                 5,941      page-faults:u                    #   44.581 K/sec                       ( +-  0.02% )
>           141,082,552      cycles:u                         #    1.059 GHz                         ( +-  0.25% )
>           158,619,823      instructions:u                   #    1.12  insn per cycle              ( +-  0.00% )
>            27,343,260      branches:u                       #  205.184 M/sec                       ( +-  0.00% )
>               743,257      branch-misses:u                  #    2.72% of all branches             ( +-  0.68% )
>    
>              0.112399 +- 0.000505 seconds time elapsed  ( +-  0.45% )
>  - 8343086: [BACKOUT] ...

There is an issue from my last push. It includes unintended merged changes from head. It's probably because I did a cherrypick for my fix when creating the initial PR. Let me close this PR and create a new one. I'll link all existing comment threads from this PR to the new PR.

-------------

PR Comment: https://git.openjdk.org/jdk/pull/21672#issuecomment-2438788544